After the iOS14 update, we’re all driving ourselves crazy trying to find audiences that really work for our brands…😓
Lookalikes based on custom events have been affected. Facebook is no longer able to track people as it did before due to the increase of opt-outs.
Of course, these changes are affecting attribution and reporting, but not the deliverability of the ads. Therefore we need to focus on the quality of the whole setup of our ads, looking into audiences, placements, and creatives.
Read More: Don’t Trust Your Facebook Rep, Always Trust the Testing Data
Today, I want to share a recent test I’ve been doing in some of our accounts and it’s been effective so far: targeting Android Devices Only.
If you haven’t tested it yet, it’s something you definitely need to consider as you might be able to find some stability in your ad account!
Here’s how I set it up:
Since you’re just targeting Android users, your audiences will be smaller so it’s important to select your best-performing countries where you know the ads will be delivered efficiently. If you are targeting worldwide and it’s performing well, then continue targeting Worldwide.
Choose your best performing 3-5 audiences and place them in a CBO with a minimum ad spend per day.
You can even split them by campaign so that you are targeting one interest or lookalike at a time.
It’s important to test cold traffic audiences first, as you will have more data than on retargeting.
Select all platforms so that the ads get more deliverability.
Choose your best 3 best post ids that have quite a lot of social proof for better results. You can also use your 3 best creatives if you still don’t have enough social proof on your ads.
I have tested this in several accounts and in most cases, I’ve been able to scale the campaigns on the first 5-7 days.
For scaling, I recommend slowly raising the budget 20% when hitting the KPIs, instead of duplicating the campaign or ad set, as the audience might not be big enough.
*You can also run an A/B test targeting Android Users vs All Mobile Devices, leave it for approximately one week and analyze the results to decide if it’s worth testing it with more audiences.